PyTorch pytorch.org

PyTorch 2.12.0 Release

Imported from official source

May 13, 2026 17:38 Software release

AI Classified by Officially

# PyTorch 2.12.0 Release Notes - [Highlights](#highlights) - [Backwards Incompatible Changes](#backwards-incompatible-changes) - [Deprecations](#deprecations) - [New Features](#new-features) - [Improvements](#improvements) - [Bug fixes](#bug-fixes) - [Performance](#performance) - [Documentation](#documentation) - [Developers](#developers) - [Security](#security) # Highlights Batched linalg.eigh on CUDA is up to 100x faster due to updated cuSolver backend selection. New torch.accelerator.Graph API unifies graph capture and replay across CUDA, XPU, and out-of-tree backends. torch.export.save now supports Microscaling (MX) quantization formats, enabling full export of aggressively compressed models. Adagrad now supports fused=True, joining Adam, AdamW, and SGD with a single-kernel optimizer implementation. torch.cond control flow can now be captured and replayed inside CUDA Graphs. ROCm

This is an extract. The publication continues at the source.

Read the original at the source: https://github.com/pytorch/pytorch/releases/tag/v2.12.0

Officially imported this from PyTorch’s own source and shows an extract. If you work there, claiming the profile and verifying the domain lets you choose to show the full text here.

This publication has changed since it was first published

2 versions recorded. The original is kept in full — nothing is overwritten.

v2 imported change on September 17, 2026 21:30 current
v1 as first published on September 15, 2026 19:08

Provenance

Organization: PyTorch — imported from official source
Official source: https://api.github.com/repos/pytorch/pytorch/releases?per_page=25 API
Imported: September 15, 2026 19:08
Versions: 2 recorded
Identity: 321486546